This is a follow-up to my recent piece “AI Coach or AI Ghostwriter? The Choice Is Yours,” which argued that AI can either sharpen your thinking or replace it. That piece was about writing. This one is about the other side of the coin: reading. The practical question is: how do you use AI to become a more productive reader rather than a lazier one?
Back in 2006, my UW students and I coined the term “machine reading” to describe the autonomous understanding of text by computers (Etzioni, Banko, & Cafarella, AAAI 2006). Two decades later, large language models (LLMs) can digest, summarize, and answer questions about text with startling competence. The irony is that the biggest consumers of this capability are people, using AI to do our reading for us.
AI-assisted reading has become so pervasive that we are approaching the absurdity captured in Tom Fishburne’s famous Marketoonist cartoon: “AI Written, AI Read.” One AI writes the memo, another AI summarizes it, with minimal human involvement.
The simplest use of AI for reading is summarization, and it certainly has its merits. Drop a 50-page PDF into your favorite LLM, ask for a summary, and you’ll get one in seconds.
But that summary is merely a skeleton. It strips away the voice, the best lines, the telling details, and the nuances that can make or break your understanding. If you are reading a legal contract, the details are the whole point. If you are reading a competitor’s product announcement, the spin they put on the numbers matters more than the numbers themselves. A skeleton doesn’t have a pulse!
AI-assisted reading punishes passivity. A recent Wharton study of over 10,000 participants found that people who relied on AI-generated summaries showed shallower knowledge and offered fewer concrete facts afterward compared to those who engaged with original sources. Advice written after AI use was shorter, less factual, and more homogeneous across users. In other words, AI summaries do not just compress text. They flatten it. Speed reading via AI can be a bit like speed dating: you cover a lot of ground, but you do not actually know anyone when you leave.
The fundamental question here is not productivity. It is about the impact of AI reading on you as the reader: What happens to your retention, your understanding, your ability to synthesize across sources? Are you winning by doing this, or are you atrophying the cognitive muscle that makes you good at your job? Outsourcing your thinking to AI is not productivity gain; it’s a competence leak.
My practical advice is: treat the summary as a triage tool, not a destination. Use it to decide whether a document deserves your time. That is genuinely valuable. The world produces more text than any human can process, and AI can help you sort the wheat from the chaff in minutes instead of hours. But once you decide that something matters, put down the summary and engage with the source.
The real power of AI reading lies not in one-shot summarization but in dialog. Think of it as an interrogation of the document, focused on what interests you. Upload the contract, the research paper, or the earnings call transcript, and then start asking questions. What are the three riskiest clauses? How does this methodology compare to the Chen et al. paper from last year? Where does the CFO’s commentary contradict the numbers in Table 4? This is not a command you fire off and forget. It is a back-and-forth conversation between you and the AI about the text, one that surfaces specific quotes, draws connections to related materials, and drills into exactly what you need. The quality of the conversation depends entirely on the quality of your questions. AI-assisted reading rewards curiosity.
A word of caution that I can’t repeat often enough: always verify anything important yourself. AI models hallucinate. They fabricate quotes, invent statistics, and present fiction with the serene confidence of a tenured professor. The verification step is essential. If you skip it, you are not reading with AI. You are gambling with AI.
You also want to adopt different reading strategies for different tasks, just as you would without AI. Summarization is fine for getting the gist of a piece, for sorting your inbox, for deciding what to read next. It will not serve you well if you need to retain the content, defend it in a meeting, or build on it in your own work. For those tasks, you need the interrogation approach, and you need to supplement it with old-fashioned human reading of the passages the AI points you to.
Used well, AI can make you a better, faster, more thorough reader by helping you navigate more material, ask sharper questions, and spot connections you would have missed. Used badly, it turns you into a consumer of predigested pablum, the intellectual equivalent of living on protein shakes when there is a farmers’ market across the street.
The machines are happy to read for you, but they won’t understand for you. The choice, as always, is yours.
Editor’s note: GeekWire publishes guest opinions to foster informed discussion and highlight a diversity of perspectives on issues shaping the tech and startup community. If you’re interested in submitting a guest column, email us at [email protected]. Submissions are reviewed by our editorial team for relevance and editorial standards.
